By Authority OfTHE UNITED STATES OF AMERICA

Legally Binding Document

By the Authority Vested By Part 5 of the United States Code § 552(a) and Part 1 of the Code of Regulations § 51 the attached document has been duly INCORPORATED BY REFERENCE and shall be considered legally binding upon all citizens and residents of the United States of America. HEED THIS NOTICE: Criminal penalties may apply for noncompliance.

Official Incorporator:THE EXECUTIVE DIRECTOROFFICE OF THE FEDERAL REGISTERWASHINGTON, D.C.

Document Name:

CFR Section(s):

Standards Body:

e

i

INTERNAL CORROSION DIRECT ASSESSMENT OF GASTRANSMISSION PIPELINES - METHODOLOGY

Final ReportGTI Contract No. 8329

January 31, 2001 – April 1, 2002

Prepared by

Oliver MoghissiCC Technologies6141 Avery RoadDublin, OH 43016

Phil DusekGas Technology Institute1700 S Mt Prospect RoadDesPlaines, IL 60018-1804

Narasi SridharSouthwest Research Institute6220 Culebra RoadSan Antonio, TX 78238

Lee NorrisScandPower, Inc.11490 Westheimer Road, No. 610Houston, TX 77077

Bruce CookinghamEl Paso Pipeline GroupNine Greenway PlazaHouston, TX 77046

Prepared for

Gas Technology Institute

1700 South Mount Prospect RoadDes Plaines, IL 60018

April 1, 2002

Final Report

ii

GRI DISCLAIMER

This report was prepared by Southwest Research Institute (SwRI) as an account ofcontracted work sponsored by the Gas Research Institute (GRI) under Contract No. 8329.Neither SwRI, GRI, members of these companies, nor any person acting on their behalf:

a. Makes any warranty or representation, expressed or implied, with respect to theaccuracy, completeness, or usefulness of the information contained in this report,or that the use of any apparatus, methods, or process disclosed in this report maynot infringe upon privately owned rights; or

b. Assumes any liability with respect to the use of, or for damages resulting from theuse of, any information, apparatus, method, or process disclosed in this report.

ACKNOWLEDGMENTSThe authors thank Keith Leewis (GTI) for his technical contributions to the work and forprogrammatic assistance. The authors also thank the project advisory group composed ofDave Aguiar (PG&E), Mike Brockman and Lee Jones (El Paso), Dave Berger (Keyspan),Garry Matocha and John Schmidt (Duke), David McQuilling and Jerry Rau (CMS), EdOndak and Richard Lopez (OPS), and Laurie Perry (SoCal). The reviews and criticalcomments provided by Sean Brossia (SwRI) and Budhi Sagar (SwRI) are also appreciated.

Final Report

iii

DEDICATION

This report is dedicated to the memory of the late Phil Dusek who initiated and sustained theproject activities.

Final Report

iv

GLOSSARY OF TERMS

Direct Assessment: a structured process for pipeline operators to assess the integrity ofburied pipelinesDirect Examination: Physical examination of a pipeline surface. For internal corrosion, thisrequires entry to the pipe.Electrolyte: A non-metallic substance (liquid or solid) that conducts electricity throughmovement of ions, thereby supporting corrosion.Fluid: A substance that does not permanently resist distortion. Both liquids and gases arefluids.

Gas Storage System: Piping and related facilities to inject and recover natural gas in anunderground formation. Recovered gas usually contains water carried from the storagestructure.Gathering System: Piping and related facilities to progressively commingle produced gasstarting from individual wells to a trunk, common, or main line. Produced gas isunprocessed.Indirect Examination: Use of tools to indirectly examine a pipeline. This includesmonitoring (e.g., sampling, coupons/probes) and inspection methods (e.g., ultrasonics,radiography, in-line inspection).Internal Corrosion Direct Assessment: An assessment methodology to determine ifinternal corrosion damage is likely or unlikely in a segment of pipe.Liquid: A substance that tends to maintain a fixed volume but not a fixed shape.

Liquid Holdup: Accumulation of liquid (i.e., input liquid volume is greater than outputliquid volume).Multiphase Flow: Flow involving more than one phase. With respect to ICDA, it refers tonatural gas and liquid water.Superficial Gas Velocity: The volumetric flow rate of gas (at system temperature andpressure) multiplied by the cross-sectional area of the pipe.Standard Cubic Feet and MMscf: Volume of gas under standard 1-atm and 60oFconditions. MMscf is million standard cubic feet.

Tariff Quality Gas: Natural gas transported by an Interstate Gas Pipeline. Tariffrequirements differ between companies, but usually include requirements for water, H2S,total sulfur, CO2, heating value, and temperature.Transmission Pipeline : A pipeline used to transport tariff quality gas over large distances.With respect to ICDA, it includes specification of nominally dry gas (e.g., Water less than7lb/MMscf).

Final Report

v

TABLE OF CONTENTS

GRI Disclaimer ....................................................................................................................ii

Acknowledgments................................................................................................................iiDedication...........................................................................................................................iiiGlossary of Terms ...............................................................................................................iv

Table of Contents.................................................................................................................v1 Project Summary......................................................................................................... 1

2 Introduction................................................................................................................. 42.1 Prediction of Corrosivity......................................................................................... 4

2.1.1 Gas Composition............................................................................................. 4

2.1.2 Water Chemistry............................................................................................. 62.1.3 Microbial Influence......................................................................................... 6

2.1.4 Velocity and Flow Effects............................................................................... 72.2 Corrosion Monitoring.............................................................................................. 72.3 Inspection or NDE................................................................................................... 8

3 Background ................................................................................................................. 93.1 Liquid Water Upsets................................................................................................ 9

3.2 Wet Gas – Condensed Water ................................................................................ 103.3 Glycol.................................................................................................................... 113.4 Other Sources of Electrolyte ................................................................................. 11

3.5 Use of Drips .......................................................................................................... 124 ICDA Method............................................................................................................ 13

4.1 ICDA in Overall Risk Management Process......................................................... 154.2 Use of Flow Modeling to Predict Liquid Accumulation Points............................ 164.3 Results of Flow Modeling..................................................................................... 18

4.4 Utilizing the Results of Flow Modeling................................................................ 224.5 Procedure for choosing detailed examination/inspection locations ...................... 25

4.6 Data Requirements for ICDA Method .................................................................. 265 Summary and Conclusions........................................................................................ 29

5.1 Validation.............................................................................................................. 29

5.2 Future Improvements ............................................................................................ 29References ......................................................................................................................... 30

Final Report

vi

LIST OF FIGURES

Figure 1. DOT/OPS year 2000 reported incidents. ............................................................. 4

Figure 2. Gas quality specifications from survey of 70 companies.1…………………….. 5Figure 3. Water content specification from survey of 70 companies................................ 10Figure 4. Deliquescence humidity for various salts after Greenspan................................ 11

Figure 5. Flow diagram of ICDA for determined length of pipe. Also consider that otherpipeline components (e.g., drips) may collect liquids............................................... 15

Figure 6. Role of ICDA in overall risk management process. .......................................... 16Figure 7. Example flow regime map for 24-inch I.D. horizontal pipe after Taitel.33 ....... 17Figure 8. Shear stress balances gravity to determine liquid holdup.................................. 18

Figure 9. Critical angles for water accumulation. For large angles and small velocities,water accumulates. For small angles and large velocities, water carries through..... 19

Figure 10. Critical Angles for water accumulation calculated by multiphase flowmodeling. ................................................................................................................... 20

Figure 11. Critical angles for water accumulation calculated by multiphase flowmodeling. Plot illustrates effect of temperature and pipe diameter........................... 20

Figure 12. F factor versus critical angle for water accumulation. Average values +standard deviation. .................................................................................................... 22

Figure 13. Screen Capture of ExcelTM spreadsheet that utilizes a Froude number, F, topredict critical inclinations for water accumulation versus gas velocity. ................. 23

Figure 14. Screen Capture of ExcelTM spreadsheet that utilizes a Froude number, F, topredict critical inclinations for water accumulation versus gas throughput.............. 24

Figure 15. Example of pipeline elevation profile and calculated inclination.................... 24Figure 16. Flow diagram of ICDA Procedure. The number represented by ‘k’ will be

adjusted based on validation of the procedure and future experience....................... 27

LIST OF TABLES

Table 1. Data Required to Use ICDA Methodology......................................................... 28

Final Report

1

1 PROJECT SUMMARY

TITLE: Internal Corrosion Direct Assessment of Gas TransmissionPipelines

CONTRACTOR: Southwest Research Institute, ScandPower, CC Technologies

PRINCIPALINVESTIGATOR:

Oliver Moghissi

REPORTPERIOD:

January 2001 to April 2002

MISSION: ‘Utilize existing technologies to develop an internal corrosionassessment methodology applied to gas transmission systems anddetermine its effectiveness.’ This report covers methodologydevelopment only.

OBJECTIVE: ‘Document and validate an assessment methodology to determineif internal corrosion is likely or unlikely to occur.’ This reportdoes not cover validation.

TECHNICALPERSPECTIVE:

Under normal operating conditions, gas transmission pipelinesare not expected to internally corrode because an upstream gasdehydration treatment facility removes the water necessary forcorrosion. The resulting gas is specified to be under-saturatedwith respect to water throughout the entire pipeline route. It isassumed that no other possibly corrosive liquids are carried overinto the gas transmission pipeline.Internal corrosion in gas transmission pipeline systems typicallyoccurs when the upstream gas processing facility delivers productthat does not meet quality specifications, since only then is itpossible for liquid (i.e., ‘free’) water (and/or other possiblycorrosive liquids) to enter the downstream transmission pipeline.Based on industry experience of gas plant operations, suchdeliveries and upsets have occurred, and in some cases havecaused internal corrosion failures.

The success of an internal corrosion control program for gastransmission pipelines depends on 1) predicting susceptibility tointernal corrosion under the full range of operating conditions,and 2) implementation of appropriate mitigation, monitoring, and

Final Report

2

inspection programs.The first issue to consider in predicting internal corrosionsusceptibility for a gas transmission pipeline is the possibility thatthe delivery of wet gas can occur in association with either of thefollowing two scenarios:

• Occasional short-term upsets at the upstream processingfacility, and/or

• Long-term, undetected delivery of gas that does not meetquality specifications.

Given either of the above two scenarios, the likelihood ofcorrosion occurring along a gas transmission pipeline depends on:

• The length of time that gas not meeting specifications isdelivered,

• The gas composition, water chemistry, microbial activity, anyother corrosive liquids associated with that gas, and

• The pipeline configuration and operating conditions resultingin local accumulation of water and/or other corrosive liquid.

Locating internally corroded pipe is difficult because the inside ofthe pipe is not easily accessible. Most existing detection methodsrequire access to the inside of the pipe for either visualexamination or inline inspection tools, and a large portion ofpipelines does not allow inline inspection because of mechanicalconstraints. Inspection techniques such as radiography andultrasonic transmission can measure wall thickness from theoutside of the pipe, but excavation (and sometimes cleaning) of aburied pipe is required. Even then, only a small area of pipe canbe inspected at a time. Therefore, a direct assessment of thelikelihood of internal corrosion through knowledge of relevantpipeline physical and operating conditions enhances the safeoperation of natural gas pipelines.

RESULTS: An internal corrosion assessment methodology applied to gastransmission systems was developed and is termed ‘InternalCorrosion Direct Assessment’ (ICDA).

TECHNICALAPPROACH:

The basis behind ICDA is that detailed examination of locationsalong a pipeline where an electrolyte such as water would firstaccumulate provides information about the remaining length ofpipe. The primary goal of the approach is to determine if internalcorrosion is likely or unlikely to exist in a chosen length of pipe.If the locations along a length of pipe most likely to accumulateelectrolyte have not corroded, then other locations less likely toaccumulate electrolyte may be considered free from corrosion

Final Report

3

and not require further examination.

PROJECTIMPLICATIONS:

The ICDA method can be used to focus the assessment of internalcorrosion in pipelines and help ensure pipeline integrity. Themethod is applicable for gas transmission lines that normallycarry dry gas but may suffer from short term upsets of wet gas orliquid water (or other electrolyte).

PROJECTMANAGER:

Narasi Sridhar

PROGRAMMANAGER:

Keith Leewis (GTI)

INDUSTRYADVISORS:

Bruce Cookingham (El Paso Energy) led the industry advisorygroup and was supported by:

Dave Aguiar (PG&E)Dave Berger (Keyspan)

Lee Jones, Mike Brockman (El Paso)Garry Matocha, John Schmidt (Duke)David McQuilling, Jerry Rau(CMS)

Gavin Nicoletta (NYS Public Service)Ed Ondak, Richard Lopez (OPS)

Laurie Perry (SoCal)David Richardson (Baker Petrolite)Keith Leewis (GTI)

Final Report

4

2 INTRODUCTION

Direct Assessment (DA) is a structured process for pipeline operators to assess the threat tothe integrity of buried pipelines.1 One of several historical threats to pipeline integrity is internalcorrosion as shown in Figure 1. Internal Corrosion Direct Assessment (ICDA) incorporates allexisting methods of examination available to a pipeline operator and provides a methodology tobest utilize those methods for specific applications. Direct examination of internal corrosion isimpractical for most pipelines because it involves exposing the inside of a buried pipeline forphysical measurements. Therefore, a suite of indirect examination tools in combination with aflow modeling approach is used to assess internal corrosion. Selection of tools depends on eachapplication, and they are broadly categorized as 1) prediction of corrosivity, 2) corrosionmonitoring, and 3) inspection or nondestructive examination (NDE). These three categories canalso be described as 1) determining if corrosion will occur in the future, 2) finding on-goingcorrosion, and 3) measuring damage that has already occurred.

DOT/OPS Reported Hazardous Liquid and Gas Transmission Incidents

Const/Mat Defect11%

Internal Corrosion12%

External Corrosion16%

Outside Force25%

Other36%

Figure 1. DOT/OPS year 2000 reported incidents.

2.1 Prediction of CorrosivityThe corrosivity of the environment inside a pipeline can be predicted based on gas

composition, water chemistry, bacteria, and velocity effects. Without water or other electrolyte,no corrosion can occur. In addition, rate determining factors act interdependently, so anycorrosion prediction that does not include them has limited accuracy and applicability.

2.1.1 Gas CompositionFor gas transmission lines, the gases considered to affect corrosion are carbon dioxide

(CO2), hydrogen sulfide (H2S), and sometimes oxygen (O2). The amount of gas in a system is

Final Report

5

defined by its partial pressure, which is the product of the total system pressure and the molefraction in the gas phase. For example, a system at 1000 psi containing 2% CO2, has a CO2partial pressure of 20 psi.

The most common form of corrosion arises from the presence of CO2 (sweet corrosion).Industry rules of thumb2 for CO2 corrosion are that 1) a partial pressure above 30 psi usuallyindicates corrosion; 2) a partial pressure between 3 and 30 psi may indication corrosion; and 3) apartial pressure below 3 psi generally is considered non-corrosive. The origins of these criteriaappear to originate from experience (i.e., anecdotal evidence) rather than theoretical prediction orcontrolled experiments. An often used correlation for CO2 corrosion is by DeWaard.3

Attack from H2S, is referred to as sour corrosion. H2S forms an acid when dissolved inwater which can accelerate corrosion, but in some cases iron sulfide deposits may reducecorrosion. However, the protection by iron sulfide is unreliable and certain forms of iron sulfidesare known to accelerate corrosion if electrolyte penetrates the corrosion product film. A relatedproblem is sulfide stress cracking (SSC), which may occur at H2S partial pressures of 0.05 psi orhigher.4

Presence of as little as 100 ppm by volume of oxygen can increase the corrosion rate inthe presence of CO2 and H2S. When O2 is present along with H2S, localized corrosion can occur,especially near the liquid-vapor interface. Oxygen increases the corrosion rate by increasing thecorrosion potential. The corrosion rate is also dependent on the pH of the water and low rates areobserved at pH values above about 6.

Gas quality specifications for gas concentration vary between companies, and waivers are oftengranted. The results of a 70 company survey are shown in Figure 2. This data is public domaininformation and thus easily available. Most of the survey information was gathered fromelectronic versions of tariffs available on the Federal Energy Regulatory Commission BulletinBoard Service (FERC BBS), while a few specific companies contributed information directly. It

should be noted that even Tariff Quality gas may have sufficient contaminants for corrosion tooccur. In addition, CO2, H2S, and O2 act interdependently because their relative concentrationsaffect the character of corrosion products. Therefore, a complete assessment of how the gascomposition affects corrosion requires consideration of all three gases and their relativeconcentrations.5

H2S, ppmv

451%

4.813%

84%

1631%

481%

CO2, volume %

15%

241%3

54%

O2, volume %

0.00122%

0.0052%

0.246%

0.44%

124%

32%

Figure 2. Gas quality specifications from survey of 70 companies.1

Final Report

6

2.1.2 Water ChemistryWater associated with natural gas is usually condensed from the gas phase or produced

from a formation. Condensed water does not contain dissolved solids. Therefore, only dissolvedgas needs to be considered (unless solids on the pipe wall are dissolved). Carbon dioxidedissolves in water and creates acidity. 6 The proton, bicarbonate ion, and undissociated carbonicacid molecule all serve as cathodic reactants for corrosion. 7

CO2 + H2O ? H2CO3 ? HCO3- + H+ ? CO3

2- + 2H+

Hydrogen sulfide similarly dissolves in water affecting corrosivity by the following dissociationH2S ? HS- + H+ ? S2- + 2H+

Produced water typically contains significant dissolved solids including sodium, iron,manganese, barium, calcium, magnesium, chloride, bicarbonate, carbonate, and sulfate.8 Thesespecies affect the deposit of solids on a pipe wall. In addition, bicarbonate and carbonate serve aspH buffers. Typical deposits are calcium and iron carbonates, barium and calcium sulfates,magnesium hydroxides and carbonates. Chloride (and other halogens) are particularly importantbecause they can cause breakdown of protective films on metals resulting in localized corrosion.

2.1.3 Microbial InfluenceMicroorganisms (primarily bacteria, but may include fungi, algae, and protozoa) can

influence corrosion of pipelines and is termed microbiologically influenced corrosion (MIC).Bacteria tend to form colonies of more than one type of microbe. These multi-memberedcolonies are also termed a community or consortium, and often form on the metal surfaces ofpipelines and other associated equipment. Biofilms can also trap solids entrained in the gasstream forming deposits that may also influence corrosion. Due to the metabolic activities ofmicrobial communities, the interface between the metal surface and the organisms may bephysically and chemically altered. Microorganisms can form acids, alcohols, ammonia, carbondioxide, hydrogen sulfide and other metabolic products capable of causing corrosion underappropriate conditions. Microbes can consume oxygen, concentrate corrosive anions (sulfatesand chlorides) in pits or crevices and under deposits, break down passive surface films, andaccelerate corrosive attack by a variety of mechanisms.9

2.1.3.1 BacteriaBacteria can live and reproduce in many different environments. For example, they can

live in acidic (low pH), neutral, or alkaline (high pH) environments, and exist over a wide rangeof temperatures and pressures. Bacteria can be introduced into oil and gas systems through avariety of mechanisms: accidental ingress of produced water, proximity of storage fields, andinadequate removal of hydrotest water. While microorganisms are found in many pipelinesystems, their presence does not necessarily cause accelerated corrosion.

Bacteria are simple one-celled organisms without true nuclei, and as a result are classifiedas among the simplest living organisms (prokaryotes). Bacteria consume nutrients from theenvironment, derive energy for life, and excrete waste products from these processes. There aremany different types of bacteria with different preferences for nutrients, temperatures, pressuresand environments. Bacteria can be classified based on their oxygen requirements. Aerobicbacteria require air or oxygen to live. Anaerobic bacteria require an environment without air oroxygen. The presence of air or oxygen can kill or inhibit the growth of some anaerobes. Obligatebacteria can only exist in one environment (either aerobic or anaerobic conditions, but not both).

Final Report

7

The growth of facultative bacteria is not restricted, and these bacteria can live in either aerobic oranaerobic environments. There are additional terms that are used to describe bacteria that are notbased on oxygen dependency. Bacteria that are attached to a surface are classified as sessile.Free-floating bacteria suspended in a fluid are classified as planktonic. Different types of bacteriacan co-exist on a metal surface in the form of a interdependent consortium. The types of bacteriatypically found in pipelines include:• Sulfate reducing bacteria (SRB) - SRB reduce sulfate to sulfide as a source of energy. Most

species of SRB contain the enzyme hydrogenase, which further catalyzes the reduction ofsulfates by hydrogen. Since iron is generally used in SRB cell structure, an iron-richenvironment often promotes their growth. Reduced sulfate reacts with available hydrogenand iron to form hydrogen sulfide and iron sulfide. SRB are strictly anaerobic, but may existin oxygen-rich environments if oxygen scavengers (e.g., aerobes, facultative anaerobes,slime-forming bacteria) are present to create locally anaerobic conditions.

• Acid producing bacteria (APB) – As a result of their metabolism (i.e., breaking down oforganic nutrients), APB can release aggressive metabolites such as organic acids, or, undercertain specialized conditions, inorganic acids. In addition to acids, APB produce alcohols,hydrogen, carbon dioxide and other metabolites that can serve as nutrients for SRB and othersimilar organisms.

• Other Bacteria - Additional types of bacteria may be present in pipelines and associatedfacilities. Facultative anaerobes and slime-forming bacteria, among others, are of mostrelevance. These organisms are important because they provide food to other communitymembers, sweep away rate-limiting or growth-inhibiting metabolites, and protect themthrough slime and/or deposit formation. These organisms can also scavenge any oxygenpresent, which can allow SRB and other strict anaerobes to exist

2.1.4 Velocity and Flow EffectsThe effect of fluid velocity on corrosion has been extensively studied (e.g., NACE10). The

majority of work has been to assess the effects of shear stress and mass transfer on corrosionrates. It should be noted that the effects of velocity are complex and depend on the metal andenvironment (e.g., Fontana 11). Many systems have no dependence on flow, many systems havehigher corrosion rates with increased flow (e.g., convection of cathodic reactants), some metalscan passivate with increased flow, and protective films are affected by a complex combination ofgrowth, dissolution, and disruption processes. The flow regime typically present in gastransmission lines and the resultant effect on corrosion are discuseed in section 4.2.

2.2 Corrosion MonitoringCorrosivity can be detected by monitoring tools, but their utility is limited by the ability to

locate them in either representative and/or the most susceptible locations along the pipeline. Forgas transmission lines, installing a monitoring coupon or probe without consideration for thelocation of most susceptible corrosion has limited use. The commonly installed coupon at theend of a line (or at a riser) can be used to identify line-wide problems in a pipeline, but thesetools are not effective for monitoring corrosion that may occur at isolated locations.

The simplest, oldest, and most widely used method to estimate corrosion is weight lossmeasurement of test coupons. A weighed sample coupon of the material under consideration isexposed to an environment and retrieved after a reasonable time interval. After removal of alldeposits, the coupon is weighed again. The weight loss is then converted to a corrosion rate. The

Final Report

8

technique requires no complex equipment or procedures, merely an appropriately preparedcoupon and a reliable means of removing corrosion product without disruption of the metalsubstrate. Most present analyses of exposed coupons include pit depth measurement andqualitative assessment of corrosion through comparison with visual (photograph) standards. Thisapproach provides information about nonuniform corrosion or pitting. Advances have been madein extended analysis and interpretation of coupon data (e.g., Eckert12).

Electrical resistance (ER) probes operate on the same principle as weight loss coupons,except the metal loss is measured through decreased electronic resistance through a wire, foil, orother thin metal structure. The advantages are that the readings are continuous allowing rapidproblem identification and the probe does not need to be retrieved (sensing is remote).

Electrochemical probes include linear polarization resistance (LPR) and electrical noise(EN). Electrochemical probes are useful for real-time measurement of corrosivity, but they mustbe located in an electrolyte to provide a reading. In addition, episodic hydrocarbon wetting canfoul the probe. Their use in dry gas pipelines (or crude pipeline) is therefore limited. LPR probesmeasure polarization resistance of a steel sample, assume electrochemical parameters (i.e., Tafelslopes), and predict corrosion rate (e.g., French13). EN probes measure small spontaneousdifferences in current and potential between nominally identical coupons (e.g., Kearns 14) toidentify the onset of localized or uniform corrosion.

Another method to monitor corrosion is to collect samples of liquid, solids, or sludgefrom a pipeline. The presence of iron may indicate corrosion. For systems where iron mayalready exist, manganese can be used to detect pipe wall dissolution. One limitation to thistechnique is that it cannot be determined if a small acceptable amount of corrosion has occurredon a large area of pipe or if a high rate of corrosion has occurred over small areas.

It should be noted that it is possible to use NDE methods for corrosion monitoring. Forexample, permanently mounting an ultrasonic tool to a pipe wall allows the inspection tool tomeasure wall loss at that location continuously over time.

2.3 Inspection or NDEPipelines are most commonly inspected for corrosion using magnetic flux leakage (MFL)

inline inspection pigs. However, a large portion of gas transmission pipelines has mechanicalconstraints preventing their use, and the tools cannot access some of the areas most susceptible tointernal corrosion (e.g., drips and stub-ends). An inspection technique such as radiography andultrasonic transmission can measure wall thickness from the outside of the pipe, but excavationof a buried pipe is required and only a small area of pipe can be inspected at a time. Othermethods include ultrasonic transmission (UT) pigs and caliper pigs. Camera inspections arepossible if the inside of the pipe is clean and can be accessed, but they are based on subjectiveinterpretation of the inspection video. Camera inspection can detect the presence of corrosionproducts and possibly pitting; however, these results can not be quantified with respect toremaining wall thickness.

Final Report

9

3 BACKGROUNDUnder normal operating conditions, gas transmission pipelines are not expected to

internally corrode because an upstream gas dehydration treatment facility removes the waternecessary for corrosion. The resulting gas is specified to be under-saturated with respect to waterthroughout the entire pipeline route. It is assumed that no other possibly corrosive liquids arecarried over into the gas transmission pipeline. Internal corrosion in gas transmission pipelinesystems occurs when the upstream gas processing facility delivers product that does not meetquality specifications, since only then is it possible for liquid water (and/or other possiblycorrosive liquids) to enter the downstream transmission pipeline. Such upsets have occurred andin some cases caused internal corrosion damage and failures (as was shown in Figure 1).

A search of publicly available information, including open literature, industry standards,and commercial models, showed that no previous work sufficiently addresses the issue ofinternal corrosion assessment in gas transmission pipelines. Much research has been performedon corrosion in wet gas systems such as gathering lines15,16 including the use of multiphase flowmodeling, 17,18,19 and commercial corrosion prediction models have incorporated this research. Inaddition, the scope of standards from organizations such as NACE International, ASTM, andAPI do not adequately cover internal corrosion specifically applied to gas transmission pipelines.It is hoped that the internal corrosion assessment methodology presented in this report will beevaluated, adopted, and evolved by these organizations. Commercial risk assessment softwarecan also serve to disseminate assessment methodologies by incorporating the technology.

Internal corrosion has occurred in systems where the gas phase is saturated with water(e.g., gathering systems), and corrosion control programs generally exist to predict, monitor,mitigate, and inspect these systems. Corrosive liquids include condensed water from gascontaining too much water vapor and liquid water that carries over from plant upsets. In addition,gas dehydration units usually use glycol (e.g., tri-ethylene glycol) which can contain water andsupport corrosion if introduced to a pipeline. Glycol can be introduced by mist carryover or byinadvertent upsets. The most effective method to prevent corrosion in gas transmission lines is toavoid introducing wet gas, liquid water, glycol, or other electrolytes to support corrosion.However, these inputs have historically occurred in gas transmission systems, so ICDA isintended to determine the corrosion related effects of these upsets.

Locating internally corroded pipe is difficult because the inside of the pipe is not easilyaccessible. Most existing detection methods require access to the inside of the pipe for eithervisual examination or inline inspection tools, and a large portion of gas transmission pipelinesdoes not allow inline inspection because of mechanical constraints. Inspection techniques such asradiography and ultrasonic transmission can measure wall thickness from the outside of the pipe,but excavation (and sometimes cleaning) of a buried pipe is required. Even then, only a smallarea of pipe can be inspected at a time. Therefore, an assessment of the likelihood of internalcorrosion through knowledge of relevant pipeline physical and operating conditions will enhancethe safe operation of natural gas pipelines.

3.1 Liquid Water UpsetsOne mechanism for liquid water to enter a pipeline is for it to be input as short, episodic

carry-over of liquid water (and may be associated with saturated gas). Generally this liquidcontains dissolved solids and is expected to evaporate in a nominally dry gas pipeline. Even ifthis water is treated with corrosion inhibitor, the remaining solids following evaporation aredetrimental to corrosion control. Input of liquid water far upstream of a gas quality monitoring

Final Report

10

point often goes undetected. This is because the water evaporates and is diluted by a largevolume of gas. However, corrosion might have occurred during the time that it accumulatedfollowing the input, and detrimental solids might have been left behind.

3.2 Wet Gas – Condensed WaterAnother mechanism for liquid water to exist in a pipeline is to condense from wet gas.

Water content in natural gas systems is commonly stated in pounds per million standard cubicfeet (lb/MMscf). Another useful method of indicating water content is in terms of dew point,which is the temperature at which the gas becomes saturated with respect to water vapor. Watersaturated (or wet) gas is the condition where water vapor is in equilibrium with liquid water; thepartial pressure of water vapor in saturated gas equals the vapor pressure of the liquid water atthe gas temperature. Since the partial pressure is proportional to the total pressure, and the vaporpressure is independent of total pressure, the dew point for a particular gas changes withoperating pressure (i.e., the dew point temperature increases with pressure).

Typical gas quality contract specifications indicate water vapor content less than 7lb/MMCF (112 g/m3).20 The results of a 70 company survey are shown in Figure 3. ASTMstandard D1142-9521 shows the relationship between water content (lb/MMscf), dew point (F),and pressure (psi). At 7 lb/MMscf (112 g/m3) and 1000 psi (6.9 MPa) pressure, the dew point is35 F (2 C).

Figure 3. Water content specification from survey of 70 companies.22

Condensed water has low dissolved solids because they cannot be carried in the vaporphase. Subsequent evaporation of condensed water therefore leaves no solids on the pipe wallunless microbial activity has produced them. In addition, detrimental species such as chlorideions are not present. However, condensed water also has no carbonates, which serve as a pHbuffer.

It should be considered that humid gas at temperatures above the dew point might alsosupport corrosion if 1) the gas is close to the dew point or 2) hygroscopic or deliquescent solids

Water Content lb/MMscf

48%

515%

63%

774%

Final Report

11

exist in the pipeline. If salts were deposited during previous liquid water upsets, water may beabsorbed by them and support corrosion in those locations. The graph shown in Figure 4 showsthe relative humidity at which various salts will deliquesce (i.e., become liquid and supportcorrosion). Mixtures of solids may have different deliquescence points. Another factor is that thepresence of solids or biofilms can slow the rate of evaporation because water becomes trapped inthe solid matrix. Fortunately, the most likely locations for accumulation of solid materials can bepredicted because they correlate with accumulation of liquid water.

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60

Temperature, degrees C

Rel

ativ

e H

um

idit

y, %

NaCl

K2CO3

MgCl2

NaOH

KCl

Figure 4. Deliquescence humidity for various salts after Greenspan. 23

3.3 GlycolIn dry natural gas systems, glycol carry-over from a dehydration facility may lead to the

formation of corrosive water/glycol mixtures.24 A common method to remove water from naturalgas is glycol dehydration. 25 Triethylene glycol (TEG) or Diethylene glycol (DEG) is used toabsorb water in the gas stream. Countercurrent contacting of wet natural gas with TEG results indry gas, but droplets of the wet glycol solution may be entrained in the gas stream and thus becarried over from the absorber into the pipeline. The glycol with absorbed water can then supportcorrosion. Glycol in a dry gas pipeline can be problematic because, unlike water, it will notevaporate under normally dry gas conditions. TEG has a vapor pressure less than 1 mm Hg at100 C.26

3.4 Other Sources of ElectrolyteOther sources of electrolyte may exist in a pipeline that should be considered as part of an

overall risk management and corrosion control program. However, the ICDA process does not

Final Report

12

specifically address sources such as remaining hydrotest water, corrosion inhibitor carriers, ormethanol for hydrate control.

3.5 Use of DripsDrips (also called drip legs or drip logs) are intended to remove free liquids from gas

pipeline systems. A wide range exists in the use and design of drips,27 with designs ranging froma short piece of pipe tied into the pipeline to configurations that are fully incorporated to themainline. Many pipelines do not have any drips, and others have drips throughout a system. Thereason for this wide range is that although drips are intended to remove liquids, they operate atfull pressure and can corrode (and have ruptured). One factor influencing the effectiveness ofdrips is the inability to locate them where liquid is most likely to accumulate. This means thatfree liquid can exist in a pipeline, but the drip does not drain it because it is not at the samelocation as the accumulation. This problem is confirmed by the experience that liquids arepushed out of a pipe by cleaning tools (i.e., pigs) despite the existence of drips. It is also possiblefor a drip configuration to prevent pigs from being run. Although drips are often placed at lowpoints in a pipeline system, selection of the proper location depends on the ability toquantitatively predict locations of liquid accumulation (e.g., by the ICDA approach).

Final Report

13

4 ICDA METHODThe basis behind ICDA for gas transmission lines is that detailed examination of

locations along a pipeline where an electrolyte such as water first accumulates providesinformation about the remaining length of pipe. If the locations along a length of pipe most likelyto accumulate electrolyte have not corroded, then other locations less likely to accumulateelectrolyte may be considered free from corrosion and not require further examination. Simplystated:

Corrosion is most likely where water first accumulates

A flow diagram shown in Figure 5 illustrates the process. The first sub-algorithm in theprocess is to predict locations where electrolyte is likely to first accumulate. This task requiresknowledge about the multiphase flow behavior in the pipe and is valid for lengths of pipe definedby potential inputs or outputs to the pipeline. The length of pipe to be considered by the ICDAprocess does not depend on distance. Rather, the ICDA applies to any length of pipe until a newinput or output changes the potential for electrolyte entry or flow characteristics.

Development of ICDA was based on a set of pipeline characteristics that define thepipelines for which ICDA as described in this paper is appropriate. The first characteristic is thatthe transported gas is normally dry (e.g., <7 lb/MMCF (112 g/m3)), and any short upsets of watereventually vaporize into the gas phase. This condition allows short-term upstream wateraccumulation, but downstream accumulation is not expected. Under this constraint, corrosion, ifit exists, will occur at isolated locations along a pipeline. These pipelines are uninhibited, do nothave internal coatings that provide corrosion protection, and are not frequently cleaned using apig. The ranges of parameters for flow modeling include an anticipated majority of gastransmission lines and are not based on technical constraints. The bounds are: a maximumsuperficial gas velocity of 25 ft/s (7.6 m/s); pipe size from 4 to 48 inch (0.1 to 1.2 m) diameter;pressure from 500 to 1100 psi (3.4 to 7.6 MPa); and relatively constant temperature over pipelength (i.e., ambient soil temperature and up to 130 F (54 C) at compressor discharge).

It should be noted that electrolyte is necessary but insufficient for corrosion. Corrosion ispossible only in the presence of an electrolyte, and the presence of corrosion damage indicatesthat electrolyte existed at that location. The absence of corrosion does not provide informationabout liquid accumulation because the factors listed in the introduction of this paper affect boththe potential driving force for corrosion and the rate. For ICDA, liquid (i.e., ‘free’) water isconsidered to be the primary source of corrosive electrolyte, glycol and wet gas are consideredsecondary, and other sources (e.g., hydrotest water) are not considered. The ICDA user isencouraged to research historical data about a pipeline since upset conditions that influenceinternal corrosion can be brief and in some cases undetected.

Locations with the longest exposure times to accumulated water (or other electrolyte) willgenerally have the most severe corrosion damage, unless the pH is such that a protective film canform. This is because water that accumulates at more than one location in a pipeline will havesimilar composition and similar corrosion rate. Gas composition is uniform throughout the lengthof pipeline until gas input or output changes the composition. When water evaporates, itconcentrates any dissolved solids, which tends to increase corrosivity. This condition tends to

Final Report

14

make the locations most likely to accumulate electrolyte the most corrosive. Microbial activityrequires water, so it is expected to be most severe at water accumulation points.

The second sub-algorithm of Figure 5 is to perform a detailed examination on locationswhere the most likely electrolyte accumulation is predicted. This detailed examination includesall of the techniques described in the introduction of this report (i.e., prediction of corrosivity,corrosion monitoring, and inspection). For many pipelines it is expected that excavation andinspection by radiography or ultrasonic transmission will be required. It should be noted thatonce a site has been exposed, installation of a corrosion monitoring tool (e.g., coupon, probe, UTsensor) may allow an operator to increase inspection intervals and benefit from real-timemonitoring in the locations most susceptible to corrosion. Corrosion monitoring tools installed atarbitrary locations (e.g., end of line) along a pipeline should not be expected to identify isolatedcorrosivity that occurs elsewhere in the pipeline. There may also be some applications where themost cost-effective approach is to run an in-line inspection tool for a portion of pipe, and use theresults to assess the downstream internal corrosion where a pig cannot be run.

If the locations most susceptible to corrosion are determined to be free from damage, theintegrity of a large portion of pipeline mileage can be assured, and resources can be focused onpipelines where corrosion is determined to be more likely. Of course, if corrosion is found, apotential integrity problem has been identified, and the method is also considered successful.

Final Report

15

INTERNAL CORROSION DIRECT ASSESSMENT

Identify locations along pipeline where free

water (or other electrolyte) first accumulates

Evaluate each location to determine if corrosion

has occurred

Results of flow modeling

Detailed examination, most likely inspection

Stop and repeat assessment based

on interval

Corrosion Found?

yes

Loop until corrosion shown to be

unlikely

Perform mitigation as part of overall integrity management program (outside scope of direct

assessment)no

If already passed through loop once, identify additional locations

Figure 5. Flow diagram of ICDA for determined length of pipe. Also consider that otherpipeline components (e.g., drips) may collect liquids.

4.1 ICDA in Overall Risk Management ProcessICDA is a method to assess the likelihood of corrosion in a given length of pipe within a

transmission pipeline. The role of ICDA in an overall risk management process is shown inFigure 6. Activities such as corrosion mitigation or repair fall outside the scope of ICDA.However, the results of an internal corrosion assessment can be used together with estimated costand consequence information to guide maintenance decisions such as repair or corrosionmitigation.

Final Report

16

Internal

CorrosionAssessment

ConsequenceAnalysis

Risk Assessment Mitigation

External

Other Failure Mode Assessment

Pipeline Risk Management

Repair/Replacement

Other

Figure 6. Role of ICDA in overall risk management process.

4.2 Use of Flow Modeling to Predict Liquid Accumulation PointsThe ICDA method relies upon the ability to identify the locations most likely to

accumulate electrolyte. These locations are predicted using the results of pipeline multi-phaseflow modeling. OLGA-S was chosen to characterize the fluid flow behavior because it betterextrapolates to field conditions than other available simulation models and is generallyconsidered to be the state-of-the-art method for prediction of liquid hold-up.28,29 This flowmodeling method, in contrast to correlative methods, applies mechanistic analysis to the relevantmultiphase flow regime. In addition, the model has been validated through large-scale laboratoryresults and comparisons to field data over a period of almost twenty years.29,30,31 The fieldvalidation of the program was carried out through the OLGA Verification and ImprovementProgram (OVIP), a research program sponsored by more than 10 oil companies, in which bothnew simulations and previous correlative methods were compared to field data. For wet gassystems, liquid holdup was found to strongly depend on gas velocity and the angle of inclination.At low rates, the liquid holdup can increase by a factor of 100 or more as the inclination anglechanges a fraction of a degree. Other models,32 which are correlation based, do not predict thisbehavior.

For gas-liquid flow, five basic flow regimes have been identified, but only two areconsidered relevant to gas transmission pipelines. An example flow map following the approachof Taitel33 is shown in Figure 7. Smooth stratified, wavy stratified, intermittent (slug and plug),annular with dispersed liquid, and dispersed bubble are possible in gas-liquid flow. In gastransmission pipelines, the volume of liquid phase (and therefore superficial liquid velocity) isassumed to be small because normal operating conditions are single phase gas, and free liquidsexist in small volumes during episodic upsets. Intermittent flow (i.e., slugging) occurs whenliquid rates are increased, and dispersed bubble flow requires a large continuous liquid phase.

Final Report

17

Annular flow requires sufficient liquid to cover the pipe wall, but even a small amount ofdispersed liquid can be entrained in the gas phase. Therefore, stratified (i.e., film) and dispersedliquid (i.e., droplet) flow regimes are relevant to gas transmission pipelines. As can be seen fromthe generic Figure 7, stratified flow occurs over a wide range of gas velocities whenever liquidsuperficial velocity (liquid flow rate divided by pipe cross sectional area) is low. This is theprevalent condition occurring in gas transmission lines.

0.1 1 10 100

0.1

1

10

100

Stratified Wavy

Annular

Slug

Dispersed Bubble

Stratified Smooth

Liqu

id V

eloc

ity, f

t/sec

Gas Velocity, ft/sec

Figure 7. Example flow regime map for 24-inch I.D. horizontal pipe after Taitel.33

Stratified film flow is considered the primary liquid water transport mechanism, and anyliquid droplets entrained in the gas are expected to evaporate because gas transmission pipelinescarry nominally dry gas most of the time. Droplets have high surface area to volume ratio, thewater is directly exposed to the gas phase, and the velocity of the gas near the droplet is high. Allthese three factors will lead to rapid evaporation of water droplets in the gas phase. Film flow incomparison has less favorable evaporative mass transfer characteristics. Liquid on the bottom ofa pipe has less surface area to volume than when dispersed as bubbles, the gas velocity at thesurface of the liquid is lower, and it is possible that a less volatile liquid covers the waterinhibiting evaporation.

Film flow along a pipe is driven by the forces of shear stress imposed by the moving gasand gravity determined by pipe inclination. Three conditions are shown in Figure 8. A downhillpipe does not accumulate water because both gas flow and gravity move liquid downstream. Ahorizontal pipe does not accumulate water if the gas is moving because the effect of gravity iszero. However, an uphill pipe creates a condition where gravity and shear stress oppose eachother. Holdup occurs when the downstream force of gravity is larger than the shear stress effect.

The balance between gravity (causing liquid to drain backwards) and shear stressbetween gas and liquid (causing liquid to be carried forward) defines the critical angle for liquidaccumulation. The effect of pipe wall roughness (e.g., solids to increase or drag reducingcoatings) is not considered significant because the shear stress at this location is small.Inclinations greater than critical will accumulate water, and inclinations less than critical will

Final Report

18

allow water to be carried downstream until a critical inclination is reached (or the water isevaporated). For a given inclination, water inventory increases when gas velocity falls below acritical threshold. For the low liquid loading encountered in gas transmission lines, this increaseis quite dramatic. Characteristically, liquid holdup fractions will jump from less than one per centto several tens of percent over a gas velocity decrease of less than 5 percent34. Wateraccumulates preferentially in first inclination exceeding a critical threshold, and continuouswater input without evaporation will eventually load all critical inclinations with water so thatlarge water input to a line will fill the first critical inclination point and carry over to the nextcritical inclination.

Shear and Gravity drive liquid downstreamNo holdup at any gas velocity

Shear drives liquid downstreamGravity drives liquid upstream

Holdup depends on slope and gas velocity

gas

Shear drives liquid downstreamGravity neutral

Holdup only with no gas flow

liquid

gas

liquid

stagnant gas

Figure 8. Shear stress balances gravity to determine liquid holdup.

4.3 Results of Flow ModelingA series of multiphase flow simulations were run to determine the effects of pressure,

temperature, gas velocity, and pipe diameter on critical angle for water accumulation. Thebounds for this parametric study were pipeline operating pressure of 500 to 1100 psi (3.4 to 7.6Mpa), temperature of 60 to 130 F (16 to 54 C), less than 25 ft/s (7.6 m/s) superficial gas velocity,and 4 to 48 (0.51 to 1.2 m) pipe diameter. Plots of critical inclination versus flow velocityillustrate the results of flow modeling. The results of predicting critical angles for 20-inch (0.51m) pipe at 900 psi (6.2 MPa) and 60 F (16 C) are shown in Figure 9. At large angles ofinclination and low gas velocities, water accumulates in the pipe. At low angles and high gasvelocities, water carries through the pipe further downstream until it reaches an inclination ofcritical angle or evaporates.

Figure 10 shows the effect of pressure on critical angle for water accumulation. Higherpressures result in water being more easily carried downstream. For a given gas velocity, thecritical angle necessary to hold up water increases with pressure. Conversely, a given inclinationon a pipeline will hold up water at lower velocities as the pressure is increased.

Final Report

19

Figure 11 shows the effects of both pipe diameter and temperature. At larger pipediameters, liquid accumulates at lower critical angles given the same gas velocity. At highertemperatures, liquid accumulates at lower critical angles given the same gas velocity, but thiseffect is relatively small. The 130 F (54 C) upper temperature bound represents a typicalcompressor station outlet temperature, which decays according to35

)exp(arg

xTT

TT

groundedisch

ground α−=−

−(1)

where T is temperature, alpha is a proportionality constant, and x is distance down the pipeline.

0

5

10

15

20

25

0 5 10 15 20 25

Gas Velocity (ft/sec)

Cri

tical

Ang

le (d

eg)

20 inch pipe (ID = 19.25")60 F900 psig

Water Accumulates

Water Carries Through

Figure 9. Critical angles for water accumulation. For large angles and small velocities,water accumulates. For small angles and large velocities, water carries through.

Final Report

20

0

5

10

15

20

25

0 5 10 15 20 25

Gas Velocity (ft/sec)

Cri

tica

l An

gle

(d

eg)

20 inch pipe (ID = 19.25")

500 Psig

700 Psig900 Psig

1100 Psig

60 degF Gas

Figure 10. Critical Angles for water accumulation calculated by multiphase flow modeling.

0

2

4

6

8

10

12

14

16

0 5 10 15 20 25

Gas Velocity (ft/sec)

Cri

tica

l An

gle

(d

eg)

700 Psig20" pipe (ID = 19.25")130 degF

700 Psig20" pipe (ID = 19.25")60 degF

700 Psig48" pipe130 degF

700 Psig48" pipe60 degF

Figure 11. Critical angles for water accumulation calculated by multiphase flow modeling.Plot illustrates effect of temperature and pipe diameter.

Final Report

21

To combine the results of simulations in an expression, a modified Froude number, F,similar to Taitel and Dukler33 is proposed here (which represents a ratio of gravitational force toinertial stress per unit area acting on a fluid)

)sin(*** 2 θρ

ρρ

g

id

g

gl

Vdg

F−

= (2)

where ρ is density, g is gravity, V is superficial velocity, and θ is angle of inclination.

The results of model runs were input to the Froude number and are plotted in Figure 12.At angles less than 0.5 degrees, F is 0.33 with a standard deviation of 0.07. At angles greaterthan 2, F is 0.56 with a 0.02 standard deviation. The angles between 0.5 and 2 degrees arebelieved to be associated with laminar to turbulent transition. F is linearly interpolated in thetransition zone.

The Froude number serves to simplify calculations, and an ExcelTM spreadsheet wasprepared so that the user can input temperature, pipe diameter, pressure, and liquid density. Twoscreen captures of the spreadsheet using gas velocity and gas throughput are shown in Figure 13and Figure 14 . User inputs are pipe size, pressure, and temperature. On a second worksheet inthe workbook, the liquid density can be adjusted, and a compressibility factor, Z, used tocalculate gas density given by

nRTPV

Z = (3)

where P is pressure, V is volume, n is moles, R is the gas constant, and T is temperature. For therange of gas conditions, a default value of 0.83 is used for compressibility based on the output ofsimulations. This value is consistent with literature values.36,37

Final Report

22

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.01 0.1 1 10 100

Angle (deg)

F (d

imen

sion

less

)

F = 0.35 + 0.08

F = 0.56 + 0.02

Transition

Figure 12. F factor versus critical angle for water accumulation. Average values + standarddeviation.

4.4 Utilizing the Results of Flow ModelingFlow modeling results are used to predict the locations at which water begins to

accumulate if it is input to the pipeline. Water accumulates on uphill sections of pipe. This isbecause the shear stress and gravity forces are balanced at this point. For a short dip associatedwith a feature (e.g., road crossing), water accumulation will occur on the short uphill segmentand therefore indicates a narrow section of pipe to examine and/or inspect. The condition wherea large up-slope exists such as would be found where a pipeline rises up a hill or mountaintogether with uncertainty or variation in gas velocity makes identification of the liquidaccumulation location within the section of pipe more difficult.

Inclination is usually given in degrees or radians and defined as change in elevation. Thesine of the inclination gives change in elevation over a distance of pipe:

)distance()elevation()sin(

∆∆

≈θ (4)

An example pipeline elevation profile is shown in Figure 15 together with the resultinginclination profile calculated by

∆∆

=)distance()elevation(

arcsinθ (5)

Final Report

23

The angles of inclination are compared to the critical angle for water accumulation predicted bythe flow modeling. The first inclination angle greater than the critical angle for accumulation isthe location where water will first accumulate. This location is therefore most likely to suffercorrosion as compared to the remaining length of pipe.

Input items in red to calculate the critical angle for water holdup:

Pipe Size I.D., Inches Pressure, psi Temperature, F22 1000 60

*Based on detailed modeling results within the range of 4 to 48 inch I.D., 500 to 1100 psi, 60 to 120F, and 0 to 25 ft/s gas velocity

Zero to Ten Degrees

0

1

2

3

4

5

6

7

8

9

10

0 5 10 15 20

Gas Velocity, ft/s

Incl

inat

ion,

deg

rees

Water Accumulates

Water Carries Through

Zero to Vertical Inclination

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50

Gas Velocity, ft/s

Incl

inat

ion

, deg

rees

Water Accumulates

Water Carries Through

Figure 13. Screen Capture of ExcelTM spreadsheet that utilizes a Froude number, F, topredict critical inclinations for water accumulation versus gas velocity.

Final Report

24

Items to calculate the critical angle for water holdup input on Main Sheet:Pipe Size I.D., Inches Pressure, psi Temperature, F

22 1000 60*Based on detailed modeling results within the range of 4 to 48 inch I.D., 500 to 1100 psi, 60 to 120F, and 0 to 25 ft/s gas velocity

Zero to Vertical Inclination

0

10

20

30

40

50

60

70

80

90

0 200 400 600 800 1000

Gas Throughput, MMSCFD

Incl

inat

ion,

deg

rees

Water Accumulates

Water Carries Through

Zero to Ten Degrees

0

1

2

3

4

5

6

7

8

9

10

0 50 100 150 200 250 300 350 400

Gas Throughput, MMSCFD

Incl

inat

ion,

deg

rees

Water Accumulates

Water Carries Through

Figure 14. Screen Capture of ExcelTM spreadsheet that utilizes a Froude number, F, topredict critical inclinations for water accumulation versus gas throughput.

-0.8

-0.3

0.2

0.7

1.2

1.7

0 10,000

20,000

30,000

40,000

46,000

53,000

60,000

66,500

Distance, ft

Incl

inat

ion

, deg

rees

0

200

400

600

800

1000

1200

1400

1600

Ele

vati

on

, ft

Elevation

Inclination

Figure 15. Example of pipeline elevation profile and calculated inclination.

Final Report

25

4.5 Procedure for choosing detailed examination/inspection locationsComparison of critical angles and actual inclinations yields locations for detailed

examination/inspection. This paper discusses the selection of individual locations along apipeline, and industry experience over time will help determine the number of redundantlocations to select for sufficient confidence to identify internal corrosion. In the near term, it maybe useful to select multiple redundant sites; this number may change as more experience isgained.

For pipelines operated at constant gas velocities, the first inclination with greater than thecritical angle represents the location where water first accumulates. All upstream inclinationswith lower angles are not expected to accumulate water and therefore are not likely to corrode.All downstream locations would either not be exposed to water (since it accumulated upstreamand evaporated), or they would be exposed only after the upstream location has filled with liquidand subsequently carried over. In this case, the upstream location would have a longer exposureperiod and therefore is expected to suffer the most severe corrosion. For the case of a pipelinewhere all inclinations are less than critical angle, the angle of highest inclination is chosen torepresent the pipeline length of interest.

Most pipelines have experienced a range of gas velocity from zero to a maximum, whichcomplicates the procedure. Critically large inclinations will trap water at any velocity up to amaximum, but upstream locations with lower angles of inclination may trap water at velocitiesless than the maximum. Because of this, examination of inclinations above the critical angle canbe used to assess the integrity of downstream pipe, but the integrity of upstream pipe remainsunknown. If information exists about the period of time a pipeline has experienced velocityranges, engineering judgement can be used to determine if short velocity changes are significant.

The procedure for the ICDA approach (considering a range of gas velocities) is shown inthe flow diagram of Figure 16:

• Find the first pipe inclination greater than the largest critical angle determined by the rangeof operating conditions and the flow modeling results. If all inclinations have angle largerthan critical, choose the angle of greatest inclination along the pipeline length.

• Perform detailed examination/inspection of the target location(s). If no corrosion is found, itis concluded that downstream corrosion is unlikely. However, if a range of velocity (or otherrelevant parameter) exists so the critical angle for accumulation may be smaller at certaintimes, upstream integrity cannot be determined by examination of a downstream inclination.

• Perform detailed examination/inspection on the location(s) with highest inclination upstreamof the initial location(s). This will provide integrity information on the pipe downstream ofthe intermediate inclination point(s) and the first inclination with angle higher than themaximum critical angle.

• Along with choosing locations having inclinations above critical angle, any fixture that cantrap water (e.g., drip, valve, stub-end) serves as an examination point. Upstream water trapscan accumulate water (or other electrolyte) before it reaches an inclination greater thancritical angle; these fixtures should therefore be examined, but they do not replaceexamination of the pipe because the rate of accumulation depends on the geometry of thefixture. Ideally, water that accumulates at a location with inclination greater than criticalangle will evaporate before filling and carrying over to the next location. However, ascenario can be envisioned where a short upset with large liquid volume fills an accumulationpoint and carries over to a fixture that traps the water. This condition is acceptable if the

Final Report

26

water evaporation rate is similar because the upstream accumulation point will be exposed tothe water for a longer period of time (and therefore suffer more corrosion). However, if thetrap geometry restricts evaporation, it is possible for corrosion to be more severe inside of thedownstream trap. Therefore, traps of similar design directly downstream of a pipe inclinationwith angle greater than critical should be examined.

4.6 Data Requirements for ICDA MethodMost of the data required to use the ICDA method is commonly available to pipeline

operators and is shown in Table 1. The exception is elevation profile of the pipeline, which mustbe known to predict the locations of electrolyte hold-up. The United States Geological Survey(USGS) has generated topographical maps that were made available to commercial softwaredevelopers. Many of the software packages include major transmission pipelines on the maps,and for those pipelines not shown, they can be located by Geographical Information System(GIS) position. To estimate pipeline elevation by surface topography, constant depth of covermust be assumed. While this may be a reasonable approximation, this uncertainty should beconsidered when selecting the hold-up locations. In addition, all features affecting elevation (andnot necessarily related to surface topography) must be identified separately (e.g., river crossings,drips, road crossing, expansion joints, etc.). If high accuracy is required, onsite pipe depthmeasurements and portable global positioning system (GPS) units can be use to accuratelydetermine pipe elevation profile and location.

Final Report

27

INTERNAL CORROSION DIRECT ASSESSMENT

Identify the first k locations along pipeline with inclination greater

than critical angle

Examine/Inspect

Corrosion Found?

yes

no

Corrosion unlikely in downstream pipe

Identify k locations with highest inclination angle

less than previous

Pipe Inlet?

yes

Internal Corrosion Unlikely or Known in

Pipeline

no

Identify the next k downstream locations

along pipeline with similar inclination angle

Figure 16. Flow diagram of ICDA Procedure. The number represented by ‘k’ will beadjusted based on validation of the procedure and future experience.

Final Report

28

Table 1. Data Required to Use ICDA Methodology

CATEGORY COMMENTS

Operating History Essential: change in direction, service, removed taps, etc.

Defined Length Essential: length between inputs/outputs

Elevation Essential: topography (pipeline location + USGS data),assume constant depth of cover

Features w/ Inclination Essential: roads, rivers, drains, etc.

Diameter Essential: ID (or OD/wall thickness)

Pressure Essential: normal operating range

Flow Rates Essential: normal operating range

Temperature Essential: conservatively assume ambient

Water Dewpoint Essential: assume <7 lb/MMSCF

Type of Inputs/Outputs Essential: need to at least know all locations

Upsets Informational: nature, intermittent or chronic?

Type of Dehydration Informational: rules out glycol input

Hydrotest Frequency Informational: presence of water

Location of Leaks/Failures Informational: supports ICDA

Other IC Data Informational: supports ICDA

Final Report

29

5 SUMMARY AND CONCLUSIONS

An internal corrosion assessment methodology applied to gas transmission systems wasdeveloped and is termed ‘Internal Corrosion Direct Assessment’ (ICDA). The ICDA method canbe used to enhance the assessment of internal corrosion in pipelines and ensure pipeline integrity.The method is applicable for gas transmission lines that normally carry dry gas but may sufferfrom short term upsets of wet gas or liquid water (or other electrolyte).

The basis behind ICDA for gas transmission lines is that detailed examination oflocations along a pipeline where an electrolyte such as water first accumulates providesinformation about the remaining length of pipe. If the locations along a length of pipe most likelyto accumulate electrolyte have not corroded, then other locations less likely to accumulateelectrolyte are considered free from corrosion and do not require further examination. Simplystated, corrosion is most likely where water first accumulates.

Many pipeline operators utilize risk management plans to prioritize areas of internalcorrosion risk and take effective mitigative measures. This includes identifying areas whereinternal corrosion (or corrosivity) exists, and conversely where internal corrosion is unlikely. Thedirect assessment methodology assesses risk from internal corrosion and incorporates all existingmethods of examination available to a pipeline operator. ICDA uses flow modeling results andprovides a framework to best utilize those methods.

Strengths of the ICDA approach include that: 1) inspection (or other examination) of pipeoutside of a high consequence area (HCA) can be used to ensure integrity inside a HCA; 2) theapproach is simple and straightforward using mature technologies; 3) it can be run by on-staffcorrosion engineers; 4) it can be used to optimize existing inspections (or any other existingassessment tool) by targeting locations of likely corrosion more accurately; and 5) it canoptimize selection of corrosion monitoring tool location.

Weaknesses of the ICDA approach include that: 1) operator familiarity and diligence isrequired (much like other assessment methods), 2) the approach applies to dry gas lines withepisodic upsets (and other additional requirements), and 3) it requires complementary tools forpipelines with extensive damage.

5.1 ValidationTo determine the uncertainty from use of ICDA and validate the method, comparison

with laboratory and field data is planned.

5.2 Future ImprovementsThe ICDA process was developed for transmission pipelines carrying nominally dry gas, and

follow-on work should be performed to cover wet gas systems such as those found in gas storageand gathering systems. Storage/Gathering systems differ from transmission systems in that theytend to use smaller diameter pipe, carry gas saturated with water, have conventional corrosionmonitoring and mitigation programs, and have many potential corrosion locations throughout apipe length. Priority was placed on transmission systems because

• A method was proposed for dry gas systems to quickly assure IC integrity of large portion ofburied pipe.

Final Report

30

• Transmission lines are more likely traverse HCA's

• Maintenance/inspection operations or shutdown have greater economic/service impacts ontransmission systems. Since each transmission line tends to carry more gas, a serviceinterruption has greater consequences than a smaller pipeline, which may even have aparallel line.

The ICDA method development was not simultaneously applied to wet gas systems so that theeffort could focus resources on single problem, and lessons learned from transmission systemscould be applied to gathering and storage.

REFERENCES

1. NACE RP on Direct Assessment, 2001, Houston TX: NACE International.

2. NACE, ‘Corrosion Control in Petroleum Production,’ TPC Publication No. 5 (1979).3. C., De Waard, U. Lotz, and D. Milliams, ‘Predictive Model for CO2 Corrosion

Engineering in Wet Natural Gas Pipelines,’ CORROSION 91, NACE International,Houston, TX (1991).

4. NACE MR-01-75, Sulfide Stress Cracking Resistant Metallic Material for Oil FieldEquipment.

5. N., Sridhar, D. Dunn, A. Anderko, M. Lencka, and H. Schutt, ‘Effects of water and GasCompositions on the Internal Corrosion of Gas Pipelines – Modeling and ExperimentalStudies, CORROSION, Vol 57, #3, NACE International, Houston, TX (2001).

6. D.M. Kern, ‘The Hydration of Carbon Dioxide,’ Journal of Chemical Education, Vol.37, No. 1. p. 14, American Chemical Society (1960).

7. C. De Waard, and D. Milliams, ‘Carbonic Acid Corrosion of Steel,’ Corrosion, Vol. 31,No. 5, NACE International, Houston, TX (1975).

8. A.G., Ostroff, Introduction to Oilfield Water Technology, NACE, Houston, TX (1979).9. H. Videla, Manual of Biocorrosion, CRC Press, Boca Raton, FL (1996).

10. NACE, ‘Flow-Induced Corrosion: Fundamental Studies and Industry Experience,’Papers presented at CORROSION/90, NACE, Houston, TX (1991).

11. M.G. Fontana, Corrosion Engineering, McGraw-Hill, New York, NY (1986).12. R. Eckert, B. Cookingham, 'Field use proves program for managing internal corrosion

in wet-gas systems,' Oil and Gas Journal, Jan 21, 2002, p 48.

13. E.C. French, and P.E. Eaton, ‘A Flush Mounted Probe for Instantaneous CorrosionMeasurements in Gas Transmission Lines, CORROSION/76 Paper #41 (1976).

14. J.R. Kearns, J.R. Scully, P.R. Roberge, D.L. Reichert, and J.L. Dawson,Electrochemical Noise Measurement for Corrosion Applications, ASTM, WestConshohocken, PA (1996).

15. B. Pound, ‘Gap Analysis of the PRCI/GRI Research Program on Internal Corrosion,’

Final Report

31

GRI Contract no. 6008, Report No. SF26363.000/AOTO/1198/BPO2 (1998).

16. Shannon, D.W., N.J Olson, R.J. Robertus, D.D. Pierce, F.O. McBarron, 1988, Effect ofWater Chemistry on Internal Corrosion Rates in Offshore Natural Gas Pipelines PRCIProject PR-3-504

17. N. Bich and K. Szklarz,’ Analysis of a Gas Field Corrosion Failure at Crossfield,’Materials Performance, vol. 7 (1999).

18. D. Richardson and C. McGovern, ‘Integration of Fluid Flow Effects within a Risk-Based Pipeline Integrity Management (PIM) Process,’ Multiphase Technology (1998).

19. P.O. Gartland and J.E. Salomonsen, ‘A Pipeline Integrity Management Strategy Basedon Multiphase Fluid Flow and Corrosion Modelling,’ CORROSION/99 Paper #622NACE Internation, Houston, TX (1999).

20. H. G. Jones, ‘Gas Quality Control and Analysis,’ AGA Oper. Section Proceedings, LosAngeles, CA and Bal Harbour, FL Meetings, May 5-7 and May 19-21 (1975).

21. Standard Test Method for Water Vapor Content of Gaseous Fuels by Measurement ofDew-Point Temperature, American society for Testing and Materials, Conshohocken,PA, V. 5.06. D1142-95 (2000).

22. R. Baldwin, 1999 Data Collected from Federal Energy Regulatory Commission BulletinBoard Service (FERC BBS).

23. L. Greenspan, J. of Research of National Bureau of Standards, 81A, 1, 89-96, USDepartment of Commerce, TIC: 241138 (1977).

24. L. van Bodegom, K. van Gelder, M.K.F. Paksa, L. Van Raam, 1987, ‘Effect of Glycoland Methanol on CO2 Corrosion of Carbon Steel,’ CORROSION/87 Proceedings,NACE International, Houston, TX.

25. A. Kohl and F. Riesenfeld, Gas Purification, Gulf Publishing Co., Houston, TX (1985).

26. Kirk-Othmer Encyclopedia of Chemical Technology, 3rd Ed., (New York: John Wiley& Sons, 1993), volume 11.

27. T. G. Braga and R.G. Asperger, ‘Engineering Consideration for Corrosion Monitoringof Gas Gathering Pipeline Systems,’ CORROSION/87 Proceedings, NACEInternational, Houston, TX (1987).

28. J. Nossen, R. Shea, and J. Rasmussen, ‘New Developments in Flow Modeling andField Data Verification,’ 2nd North American Conference on Multiphase Technology,BHR Group Limited, June 21-23 (2000).

29. R. Shea, J. Rasmussen, P. Hedne, and D. Malnes, ‘Holdup predictions for wet-gaspipelines compared,’ Oil & Gas Journal, May 19 (1997).

30. K.H. Bendiksen, D. Malnes, R. Moe, and S. Nuland, ‘The Dynamic Two-Fluid ModelOLGA: Theory and Application,’ SPE Production Engineering, May 1991, p. 171.

31. C. Alvarez and M. H. Al-Awwami, ‘Wet Crude Transport Through a Complex HillyTerrain Pipeline Network,’ 74th Annual technical Conference and Exhibition, SPE56463 (Houston, TX, Society of Petroleum Engineers, October 3-6, 1999).

32. H.D. Beggs and J.P. Brill, ‘A study of Two-Phase Flow in Inclined Pipes,’ Journal of

Final Report

32

Petroleum Technology, May (1973), p. 607.

33. Y. Taitel, A. E. Dukler, ‘A Model for Predicting Flow Regime Transitions in Horizontaland Near Horizontal Gas-Liquid Flow,’ AIChE Journal, 22, 1 (1976), p47.

34. R. Shea, J. Rasmussen, P. Hedne, and D. Malnes, ‘Holdup predictions for wet-gaspipelines compared,’ Oil & Gas Journal, May 19 (1997).

35. R.N. Parkins, ‘Overview of Intergranular Stress Corrosion Cracking ResearchActivities,’ PR-232-9401, PRCi (1994).

36. R. Reid, J. Prausnitz, and T. Sherwood, The Properties of Gases and Liquids, (NewYork: McGraw-Hill , 1977).

37. R. Perry and D. Green, Perry’s Chemical Engineers’ Handbook, 6th edition, (NewYork: McGraw-Hill, 1984).